Embrapa is in the final stages of developing ProCarbon-Soil (Procs), the first model developed in a tropical region to measure the dynamics of so-called carbon farming, which are agricultural activities aimed at removing carbon from the atmosphere and storing it in the soil. The new model was presented in an article published in the Soil Science Society of America Journal. In addition to being adapted to the conditions of Brazilian agriculture, the model improves the compatibility between measured and simulated variables, which can contribute to making the carbon market more transparent for operators and farmers. While traditional multicompartmental soil carbon models use between four and eight state variables and between seven and 20 parameters—some of them complex and difficult to measure—the proposed model uses only two measurable variables: total carbon stock and its decomposability. Based on a mathematical approach and the theoretical framework on the dynamics of organic carbon in the soil, Procs accurately simulates the impacts of agricultural practices such as crop rotation, no-till farming, and soil preparation under different climatic conditions. To make the development possible, a primary database composed of research information from Embrapa and Bayer, a partner company in the project, was used. Calibrating the model required another extensive secondary database, composed of 4,290 soil samples from all Brazilian regions, originating from 370 published studies. “What we need for the carbon market is the dynamics, the variation of total carbon. A second metric is associated with the quality of this carbon, how stable it is. Other models partition carbon into different compartments. Ours evaluates the overall quality of carbon,” says Luis Gustavo Barioni, researcher at Embrapa Agricultura Digital (SP) and leader of the project. Barioni explains that, in addition to being measurable, verifiable, and reportable in a simpler way, the proposed model brings advantages to Brazilian agriculture. Until then, all other models recognized in the carbon market had been developed in temperate regions, where production systems are different and the behavior of organic matter in the soil is also distinct. “Among these advantages, autonomy and continuous development stand out, ensuring it doesn't become obsolete. It's our own model; we own the intellectual property, which is another important advantage,” says Barioni. For Procs to be adopted, it needs accreditation from certification bodies operating in the voluntary carbon market. Currently, researchers are preparing a Model Validation Report to be submitted to Verra, the world's largest carbon credit certification body. Procs was developed within the scope of the PRO Carbono project, a partnership between Embrapa and Bayer. It will be integrated into a large platform of the multinational; in addition, it will support the National Inventory of Greenhouse Gas Emissions and Mitigation and public policy actions. “This model needs to be recognized by academia and other actors. The more people use it, the greater its credibility will be,” explains Barioni.
Precision and reliability
There are two ways to measure the dynamics of carbon farming. One is based on soil samples taken at time zero and repeated after a period of time. This method is expensive, laborious, and time-consuming. The second method is through models recognized by carbon market certification bodies. In studies, Procs has demonstrated the ability to produce trajectories very similar to those generated by the Century model, a classic multicompartmental model. For a 50-year simulation period, the root mean square error between the model estimates was 1.03 tons of carbon per hectare. According to the researchers responsible, the uncertainty of Procs is even lower compared to that found in field measurements with traditional protocols. In direct measurements of carbon stock variation in typical Brazilian farm plots of approximately 40 hectares, the error reaches 3.8 tons per hectare. In addition to having an acceptable degree of precision to be recognized by certification bodies, Procs seeks to overcome challenges in carbon trading (Carbon Farming Trading Schemes – CFTS) and make carbon credit projects more viable, reliable, and auditable. Model and Data Fusion Another advantage of ProCarbon-Soil is its compatibility with new model and data fusion technologies (also called data assimilation). Through these technologies, it will be possible to automatically feed the models with satellite data, for example. And with the use of model-data fusion techniques and artificial intelligence, it will be possible to identify and correct deviations according to the measurement. “This is what has been called data learning — data assimilation and machine learning — combined with model-data fusion. Our model is better suited for this, and we believe that these new technologies will be increasingly associated with estimating the variation in carbon stocks,” explains Luis Gustavo Barioni. The article "ProCarbon-Soil: A dynamic model for improved model-data compatibility in carbon farming" was published in the May issue of the Soil Science Society of America Journal. In addition to Luis Barioni, the work is co-authored by project fellows Beatriz Valladão, Vitor Mourão, Yusuf Karatay, Junior Damian, and Vinicius Melício. Also co-authored are Robert Ewing of Climate LLC (United States), Rodrigo Rejaili of Bayer Crop Science, and Rafael Silva of the University of Edinburgh (Scotland).
This text was translated by machine from Brazilian Portuguese.